Count items in database as much as possible
#344
Merged
elegaanz
merged 6 commits from count-in-db
into master
5 years ago
Reviewers
Request review
No reviewers
Labels
Clear labels
Related to the REST API
Code running on the server
Stuff related to Federation
Related to the front-end
Translations, and related code
More about project management or code than the project itself
The building, or installation process of Plume
Something isn't working
We need to talk
New feature or request
This is a new feature
Compatibility with different browsers, readers and OS
Related to an external package that Plume uses
UI/UX related issues and PRs
Good for newcomers
Extra attention is needed
Issues affecting only mobile UX
How elements're rendered out for the end user
Something else needs to be fixed first
This issue or pull request already exists
This PR is not complete yet
Issues concern a limited number of instances
This doesn't seem right
Need to be discussed by the community (on Loomio)
This PR is ready to be reviewed
Proposed ideas worth considering
This is issue has been created after a vote on Loomio
This will not be worked on
Apply labels
A: API
Related to the REST API
A: Backend
Code running on the server
A: Federation
Stuff related to Federation
A: Front-End
Related to the front-end
A: I18N
Translations, and related code
A: Meta
More about project management or code than the project itself
A: Security
Build
The building, or installation process of Plume
C: Bug
Something isn't working
C: Discussion
We need to talk
C: Enhancement
New feature or request
C: Feature
This is a new feature
Compatibility
Compatibility with different browsers, readers and OS
Dependency
Related to an external package that Plume uses
Design
UI/UX related issues and PRs
Documentation
Good first issue
Good for newcomers
Help welcome
Extra attention is needed
Mobile
Issues affecting only mobile UX
Rendering
How elements're rendered out for the end user
S: Blocked
Something else needs to be fixed first
S: Duplicate
This issue or pull request already exists
S: Incomplete
This PR is not complete yet
S: Instance specific
Issues concern a limited number of instances
S: Invalid
This doesn't seem right
S: Needs Voting/Discussion
Need to be discussed by the community (on Loomio)
S: Ready for review
This PR is ready to be reviewed
Suggestion
Proposed ideas worth considering
S: Voted on Loomio
This is issue has been created after a vote on Loomio
S: Wontfix
This will not be worked on
No Label
A: API
A: Backend
A: Federation
A: Front-End
A: I18N
A: Meta
A: Security
Build
C: Bug
C: Discussion
C: Enhancement
C: Feature
Compatibility
Dependency
Design
Documentation
Good first issue
Help welcome
Mobile
Rendering
S: Blocked
S: Duplicate
S: Incomplete
S: Instance specific
S: Invalid
S: Needs Voting/Discussion
S: Ready for review
Suggestion
S: Voted on Loomio
S: Wontfix
Milestone
Set milestone
Clear milestone
No items
No Milestone
Assignees
Assign users
Clear assignees
No Assignees
2 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.
No due date set.
Dependencies
No dependencies set.
Reference: Plume/Plume#344
Reference in New Issue
There is no content yet.
Delete Branch 'count-in-db'
Deleting a branch is permanent. It CANNOT be undone. Continue?
No
Yes
Codecov Report
0% <0%> (ø)
12.01% <0%> (-0.13%)
0% <0%> (ø)
0% <0%> (ø)
0% <0%> (ø)
0% <0%> (ø)
0% <0%> (ø)
0% <0%> (ø)
0% <0%> (ø)
0% <0%> (ø)
Continue to review full report at Codecov.
Things I commented are further optimizations possible, however they're not strictly needed to be implemented for this to be merged. Actually I don't even know if it would give better performance as optimizations I proposed can be proved equivalent to current implementation with what DBMSs know about our model, so they might do the optimization themselves
I wonder if the second query is useful, I think
count()
ing on the first one is sufficient, except in case where :.select(blog_authors::author_id)
is removed too, if it is not then multiple (id_X,blog_1,author_1) would be mapped to only one (author_1) anyway. Hope I'm clear enough)ordering then counting is useless, I guess PostgreSQL would see it and optimize that away, but I also think SQLite is too dumb to do the same, so it might be shorter and faster to remove that
Yes, I just copy/pasted another function. I'll change that.
OK, I'll change it (and yes you were clear enough 🙂)
Love it when an issue is only reproducible on one db backend. It might be because sqlite don't enforce some constraints by default, we could enabled
PRAGMA foreign_keys = ON;
, but I wonder what changes to make this test failwe have that enabled — on the
users
table — not sure if/how we enable it by default.maybe we're finally using that constraint? 😅
@igalic I think we only set it in a migration, this property is not stored so we need to set it somehow at each Plume launch, and this is a global var, so it was enabled for all tables (but only until the migration end)
unlike (some) other
pragma
instructions, it's not storable… it's, as you say, per connection or per transaction.So perhaps the diesel sqlite driver has something we can turn on by default
i checked, it can't
but maybe we can wrap the query functions with a
begin transaction; pragma foreign_keys = on;
@ -41,0 +48,4 @@
sql_query("PRAGMA foreign_keys = on;").execute(conn)
.map(|_| ())
.map_err(|_| ConnError::ConnectionError(ConnectionError::BadConnection(String::from("PRAGMA foreign_keys = on failed"))))
}
oh, I was searching for exactly that but did not find it
@ -41,0 +48,4 @@
sql_query("PRAGMA foreign_keys = on;").execute(conn)
.map(|_| ())
.map_err(|_| ConnError::ConnectionError(ConnectionError::BadConnection(String::from("PRAGMA foreign_keys = on failed"))))
}
Looks like it is not working at all🙃 nvm I think I found the issue@ -41,0 +48,4 @@
sql_query("PRAGMA foreign_keys = on;").execute(conn)
.map(|_| ())
.map_err(|_| ConnError::ConnectionError(ConnectionError::BadConnection(String::from("PRAGMA foreign_keys = on failed"))))
}
I think your changes are only affecting the web interface, which is at the same time the most important part, and the one not tested by travis 🙃
I've tested doing similar changes in plume_model::lib:tests::db (adding
sql_query("PRAGMA foreign_keys = on;").execute(&conn).unwrap();
) which seems to work fine"bfore", "after bfore", "after after", then is it "after after bfore" or "after after after"?
(joking about the comments, but they should be removed actually, I guess you forgot to)
Oops 😬
Reviewers
38302203f4
.Step 1:
From your project repository, check out a new branch and test the changes.Step 2:
Merge the changes and update on Forgejo.